Uploaded July 2026 | Updated September 2026, 2 weeks ago
Don't miss out! Join us at our next KubeCon + CloudNativeCon events in Yokohama, Japan (29-30 July, 2026), and Shanghai, China (8-9 September, 2026) Salt Lake City, United States (Nov 9–12, 2026). Connect with our current graduated, incubating, and sandbox projects as the community gathers to further the education and advancement of cloud native computing. Learn more at kubecon.io
The Leapfrog Upgrade Playbook - Yug Gupta, Walmart Global Tech
This is a story of how Walmart Leapfrogged 11 Kubernetes Releases Without Customer Downtime.
If that sounds impossible, wait untill you see the constraints: Millions of k8s resources, an etcd database nearing ~7GB, and a version gap so wide that “just upgrade” wasn’t a plan. We needed a method that was safe across API removals, predictable under extreme object counts, and repeatable enough to become a standard operating procedure.
Attendees will learn: (1) how to preflight deprecated APIs and CRDs using upstream guidance (2) how to design restore ordering and “selective restore” boundaries to avoid bringing back broken objects, (3) how to tune API server concurrency so backup/restore doesn’t become a self-inflicted outage, and (4) how to validate correctness and plan rollbacks.
This is a field guide for anyone who’s fallen far behind and needs a safe way forward without betting production on a one-shot upgrade.
No secret sauce, just the hard-earned physics of Kubernetes at scale.
Don't miss out! Join us at our next KubeCon + CloudNativeCon events in Yokohama, Japan (29-30 July, 2026), and Shanghai, China (8-9 September, 2026) Salt Lake City, United States (Nov 9–12, 2026). Connect with our current graduated, incubating, and sandbox projects as the community gathers to further the education and advancement of cloud native computing. Learn more at kubecon.io
The Leapfrog Upgrade Playbook - Yug Gupta, Walmart Global Tech
This is a story of how Walmart Leapfrogged 11 Kubernetes Releases Without Customer Downtime.
If that sounds impossible, wait untill you see the constraints: Millions of k8s resources, an etcd database nearing ~7GB, and a version gap so wide that “just upgrade” wasn’t a plan. We needed a method that was safe across API removals, predictable under extreme object counts, and repeatable enough to become a standard operating procedure.
Attendees will learn: (1) how to preflight deprecated APIs and CRDs using upstream guidance (2) how to design restore ordering and “selective restore” boundaries to avoid bringing back broken objects, (3) how to tune API server concurrency so backup/restore doesn’t become a self-inflicted outage, and (4) how to validate correctness and plan rollbacks.
This is a field guide for anyone who’s fallen far behind and needs a safe way forward without betting production on a one-shot upgrade.
No secret sauce, just the hard-earned physics of Kubernetes at scale.
